Evaluation of Cooperative Spectrum Sensing Based on Large Scale Measurements

Dynamic spectrum access to unused licensed spectrum has been proposed as solution to the spectrum scarcity problem. Reliable detection of any primary user signal is crucial for the required interference-free operation. It has been shown that cooperative sensing has the potential to alleviate propagation effects such as fast fading or shadowing and thus to improve sensing reliability. However, there is lack of practical studies confirming the theoretical results. In this paper we present our measurement setup for cooperative sensing and detailed experimental results that prove the major findings. Cooperative sensing can improve the detection reliability and lower the sensitivity requirements on single sensors. We also validate that correlated spectrum measurements will lower the cooperation gain and that such correlations decrease with distance. Additionally, we develop a metric for the robust representation of these correlations and discuss the impact on practical system design. More precisely, we elaborate on different modelling approaches that cognitive radios may apply in order to learn about their radio environment and to exploit such knowledge for enhanced dynamic spectrum access.